Ontology Modeling: Distinguishing Parts from Kinds
CraigTrim 110000G799 Visits (3889)
Here is a simple Ontology:
If the “X is a Y” relation are logical (does the relationship make sense in the real world?), then the Ontology class hierarchy is valid.
Here's an example of an invalid Ontology class hierarchy:
This model asserts that “Male is a Father”. Or, if you were to express this in a somewhat more formal manner, “all males are fathers”. True for some, not true for others. So this is an illogical relation
There are many ways to express hierarchies.
A common biological hierarchy is:
This is a great use of a tree-like structure to demonstrate that
Kingdoms contain PhylumsLikewise, it shows that
Species are a part of Genus,This tree structure (however useful under many circumstances) is not a valid Ontology.
In an Ontology, we would have to read this as:
And none of these statements are true.
In the first Ontology model, a Father is a Person and a Mother is a Person. But a Genus is not a Species, and a Phylum is not a Kingdom.
It is important to distinguish between parts and kinds:
This is often the difference between a Taxonomy and an Ontology. A Taxonomy doesn’t need to distinguish between parts and kinds. An Ontology must make this distinction.
Note that as we abbreviate “is a kind of” to “is a”, we can likewise abbreviate “is part of” to “has a”. So again, we can say “Order has a Family” but not “Family is a Order”